THE TROPICAL 
AGfilCULTURIST. [May 1, 1901. 
grammes of cassava into 500 cubic centimetres of 
•water which was left for two hours in a wen-clofed 
vefpel. It was next strained through a linen cloth 
into a tlBsk wliic?i was then corked and allowed to 
stand until the starch subsided. 200 cubic centi- 
metres of the clear liquid were decanted and distilled 
and the results obtained wei'e mnltiiilied by four for 
the total in 500 grammes. This calculation was 
based on the assumption that the original 500 
grammes contributed 300 cubic centimetres of water 
— an assum])tion which is not far from the truth 
if the cassava is (|uite fresh. T^'rancis also remarks 
that a slight loss occurred during the I'l'ocess of 
gratinK: and in a tropical laboratory this loss might 
be arpr'ciablfs. 1 endeavoured to avoid this loss on 
the one hand, and the possibility of error through 
assuming that ail cassavas contained 60 per cent, of 
water on the other, by taking the whole root, slicing 
it quickly, and allowing it to stand in water all night 
in a well-corked flask. The liquid was decanted in 
the morning and the whole distilled. The extraction 
with Water was repeated a second and a third time. 
The slices effectually prevented the great disadvan- 
tage arising from the starch finding its way into the 
distilling flask, which Francis prevented by sub- 
sidence but which necessitated his distilling an 
aliquot part, this aliquot part being a ((('finite 
volume of 200 cubic centimetres taken from an 
approxiwatf total volume of 800 cubic centimetres. 
, Francis does not appear to have iiade more than 
one extraction with water. I liave made three in all 
cases and in some as many as five. 1 have found 
that with slices the second extraction yields as 
much, and sometimes more, than the first. I have 
repeated Francis's method of grating and find that a 
second extraction yields about half as much prussic 
acid the first. It would appear from thi.s, there- 
fore, that Francis's results are not as high as they 
would have been had he made a second extraction. 
The figures obtained by S'rancisiatid myself, are: — 
HCN per cent. 
' ' 1 
Francis. Carmody. 
Mean 0-0168 ... 0010 
Highc,*t 0-0238 ... 0019 
Lowest 0-0113 ... 0-005 
It will be seen from this that while my "lowesf 
percentage is much below that of Francis my 
"highest" is nearly the same as his. This establishes 
the truth of the statement made by Francis, "that 
the so-called sweet, or harmless cassava not only 
yielded prussic acid, but the quantity obtained from 
it BO nearly e(iualled that from the hitter that no line 
of distinction could be drawn between them," so far 
as it relates to the amount of prussic acid in the 
cassava; but I have discoveretl an analytical differ- 
ence between the two kirrds by means of which they 
can be satisfactorily distinguished. 
This discovery arose in trying to avoid the difB- 
culties caused by the presence of siarch in the distil- 
ling flask At first 1 placed the slices in the distil- 
ling flask, added water in one set of experiments, and 
passed steam through in another and collected the 
distillates. This failed in each case through the 
gelatinising action of the boiling water or steam on 
the starch. This starch being chiefly confined to the 
inner protions, in my next experiments I soaked the 
inner slices in water at ordinary temperature, decant- 
ed and distilled. The quantity of starch thus ad- 
mitted into the flask was so small as not to interfere 
with the distilla/tion. The skin and outer cortical 
layer, wliich together •constitute about one-fifth of 
the total weight of the sweet cassava, could be 
placed direct in the distilling flask, and the distilla- 
tion carried on satisfactoi-ily. It was this separation 
of the parts for convenience in distillation that led to 
the discover.v that in sweet cassava the prussic acid 
is located chiefly in the skin and outer cortical layer. 
The following are some of the results obtained 
from fairly representative samples of sweet cassava. 
Inner part 
HCN 
per cent 
0-006 
0003 
0-015 
0-008 
0-011 
0 008 
Skin and outer 
cortical layer. 
HCN 
per cent. 
0-033 
O-OU 
0-033 
0-031 
0-020 
0-032 
Inner ftart 
HCN 
per cent. 
0-004 
0-010 
o-ooi 
0-005 
0-003 
Skin and outer 
cortical layer. 
HCN 
per cent. 
0-024 
0-030 
0-042 
0-038 
0-034 
AVith bitter cassava the following results were ob- 
tained from representative samples: — 
inner part. 
HCN 
per cent. 
0-031 
0 034 
0-021 
0-037 
0-030 
0-014 
Skin and outer 
cortical layer. 
HCN 
per cent. 
0-024 
0-012* 
0-025 
0-014* 
0-025 
0 013 
Skin and outer 
Inner part, cortical layer. 
HCN 
HCN 
per cent. 
per cent 
0017 
0019 
0-019 
0-020 
0-016 
0-024 
0-017 
0 020 
0-013 
0-016 
0-032 
0-035 
"Exceptional, 
The general conclusions to be drawn from these 
results ai e: (1) that in sweet cassava the prussic 
acid is net uniformly distributed throughout the tuber 
and that in bitter cassava it is uniformly distributed, 
or nearly so; and (2) that this affords an analytical 
means of distinguishing between swet and litter 
cassava. 
Cassava is cooked either by roasting the entire 
tuber (this removes all the HCN) or by boiling. An 
important observation in connection with local 
methods of eating cassava is that the skin of the 
sweet kind is removed before boiling, the inner 
portion alone being cooked and eaten. This custom 
is no doubt the result of experience and accords with 
what science would now teach as a wise precaution. 
But, notwithstanding this removal of the skin, ill- 
effects are said to follow from the use of the inner 
portion if water is drunk some time after a meal. 
This may now be accounted for in this way. It ig 
shown in my experiments that whilst a first boiling 
removes a certain amount of hydrocyanic acid — all, 
in fact, that exists ready formed in the tuber — that a 
second addition of water and subsequent boiling 
removes a further portion. It would appear from 
this that cassava in a person's stomach would also 
develo]) an amount of prussic acid equivalent to 
that obtained in a second distillation, and that the 
drinking of water would dissolve this and cause it to 
act more readily on the system. The total quantity 
would, however, be far below the minimum fatal 
dose and would only bo expected to produce those 
unpleasant ; ssults occasionally reported. The ques- 
tion next arises — Is this second portion of hydro- 
cyanic acid produced by fermentative changes, as is 
known to happen in the case of bitter almonds; or ia 
it part of that originally existing in the tuber and in 
some way protected from the solvent action of the 
first ti-eatment with water? It ia possible that, with 
slices, the water would penetrate slowly to their in. 
terior, and thus produce a second yield of prusaig 
